A Novel Time-Incremental End-to-End Shared Neural Network with Attention-Based Feature Fusion for Multiclass Motor Imagery Recognition

<div>Our method of SCNN-BiLSTM network based on attention. The EEG signals are input to convolution layers and therefore preprocessed time series are generated. Moreover, the time series are input to BiLSTM cells for the exchange of information among different time points. And the attention mechanism module receives the output of BiLSTM cells, calculates weights for different time points, and outputs the ultimate result.</div>

Computational Intelligence and Neuroscience

fig1

Figure 1

Figure 1: A Novel Time-Incremental End-to-End Shared Neural Network with Attention-Based Feature Fusion for Multiclass Motor Imagery Recognition 